Mechanism of the alkali degradation of (6-4) photoproduct-containing DNA

Article abstract of DOI:10.1039/c2ob06966k

The (6-4) photoproduct is one of the major damaged bases produced by ultraviolet light in DNA. This lesion is known to be alkali-labile, and strand breaks occur at its sites when UV-irradiated DNA is treated with hot alkali. We have analyzed the product obtained by the alkali treatment of a dinucleoside monophosphate containing the (6-4) photoproduct, by HPLC, NMR spectroscopy, and mass spectrometry. We previously found that the N3-C4 bond of the 5′ component was hydrolyzed by a mild alkali treatment, and the present study revealed that the following reaction was the hydrolysis of the glycosidic bond at the 3′ component. The sugar moiety of this component was lost, even when a 3′-flanking nucleotide was not present. Glycosidic bond hydrolysis was also observed for a dimer and a trimer containing 5-methyl-2-pyrimidinone, which was used as an analog of the 3′ component of the (6-4) photoproduct, and its mechanism was elucidated. Finally, the alkali treatment of a tetramer, d(GT(6-4)TC), yielded 2′-deoxycytidine 5′-monophosphate, while 2′-deoxyguanosine 3′-monophosphate was not detected. This result demonstrated the hydrolysis of the glycosidic bond at the 3′ component of the (6-4) photoproduct and the subsequent strand break by β-elimination. It was also shown that the glycosidic bond at the 3′ component of the Dewar valence isomer was more alkali-labile than that of the (6-4) photoproduct. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2ob06966k

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Organic &
Biomolecular
Chemistry
Cite this: Org. Biomol. Chem., 2012, 10, 2318
www.rsc.org/obc
PAPER
Mechanism of the alkali degradation of (6–4) photoproduct-containing DNA†
Norihito Arichi,a Aki Inase,‡a Sachise Eto,b Toshimi Mizukoshi,b Junpei Yamamotoa and Shigenori Iwai*a
Received 28th November 2011, Accepted 19th December 2011
DOI: 10.1039/c2ob06966k
The (6–4) photoproduct is one of the major damaged bases produced by ultraviolet light in DNA. This
lesion is known to be alkali-labile, and strand breaks occur at its sites when UV-irradiated DNA is treated
with hot alkali. We have analyzed the product obtained by the alkali treatment of a dinucleoside
monophosphate containing the (6–4) photoproduct, by HPLC, NMR spectroscopy, and mass
spectrometry. We previously found that the N3–C4 bond of the 5′ component was hydrolyzed by a mild
alkali treatment, and the present study revealed that the following reaction was the hydrolysis of the
glycosidic bond at the 3′ component. The sugar moiety of this component was lost, even when a
3′-flanking nucleotide was not present. Glycosidic bond hydrolysis was also observed for a dimer and a
trimer containing 5-methyl-2-pyrimidinone, which was used as an analog of the 3′ component of the
(6–4) photoproduct, and its mechanism was elucidated. Finally, the alkali treatment of a tetramer,
d(GT(6–4)TC), yielded 2′-deoxycytidine 5′-monophosphate, while 2′-deoxyguanosine 3′-monophosphate
was not detected. This result demonstrated the hydrolysis of the glycosidic bond at the 3′ component of
the (6–4) photoproduct and the subsequent strand break by β-elimination. It was also shown that the
glycosidic bond at the 3′ component of the Dewar valence isomer was more alkali-labile than that of the
(6–4) photoproduct.
photoproduct is the same photocycloaddition as that for the
CPD, but it occurs between the C5–C6 double bond and the C4
Introduction
DNA is subjected to various reactions with endogenous and
exogenous factors in cells, and the alteration of its chemical
structure, referred to as DNA damage, may induce carcinogen-
esis or cell death unless cellular repair systems properly func-
tion.1 Ultraviolet (UV) light is one of the factors that induce
DNA damage, and produces two major types of lesions, the
cyclobutane pyrimidine dimer (CPD) and the pyrimidine(6–4)
pyrimidone photoproduct ((6–4) photoproduct), between two
adjacent pyrimidine bases.2 While a cyclobutane ring is formed
between the two C5–C6 double bonds by the [2 + 2] cyclo-
addition in the CPD case, the (6–4) photoproduct (1) has a
covalent bond between the C6 and C4 positions of the 5′ and 3′
components, respectively, and the oxygen/nitrogen atom at the
C4 of the 3′ thymine/cytosine is transferred to the C5 of the 5′
base.3–5 The first reaction in the formation of the (6–4)
carbonyl/imino group of the 5′ and 3′ pyrimidine bases, respect-
ively. This reaction produces an oxetane/azetidine intermediate,
and then the C4–O/C4–N bond in this intermediate is cleaved to
yield a stable structure (1).6
The (6–4) photoproduct is reportedly more mutagenic than the
CPD.7–11 A remarkable feature is that a T → C or C → T tran-
sition occurs with high frequency at the 3′ component of the
(6–4) photoproduct. Moreover, the (6–4) photoproduct blocks
replication by translesion DNA polymerases that can bypass a
CPD.12,13 This is probably due to the alterations in the chemical
structure and the orientation of the 3′ base. The intramolecular
hydrogen bond suggested in our previous study14 may also exert
some influence. To maintain genetic integrity, the (6–4) photo-
product must be repaired by the nucleotide excision repair
pathway in human cells,15,16 while some other organisms have
an enzyme, (6–4) photolyase, that directly converts this photo-
product to the original pyrimidine bases, using blue light
energy.17
From the viewpoint of chemistry, the formation of the (6–4)
photoproduct, its isomerization to the Dewar valence isomer by
exposure to longer wavelength UV light, and the light-dependent
repair by (6–4) photolyase have been studied and discussed in
detail.17–20 On the other hand, the chemical reactions of the
(6–4) photoproduct have not been investigated. UV irradiation of
DNA produces alkali-labile sites,21 and the product formed at
aDivision of Chemistry, Graduate School of Engineering Science, Osaka
University, 1-3 Machikaneyama, Toyonaka, Osaka560-8531, Japan.
E-mail: iwai@chem.es.osaka-u.ac.jp
bFundamental Technology Labs., Institute for Innovation, Ajinomoto
Co., Inc., 1-1 Suzuki-cho, Kawasaki-ku, Kawasaki, Kanagawa210-8681,
Japan
†Electronic supplementary information (ESI) available: Synthetic
procedures and characterization data. See DOI: 10.1039/c2ob06966k
‡Present address: National Cancer Center Research Institute, 5-1-1
Tsukiji, Chuo-ku, Tokyo 104-0045, Japan
2318 | Org. Biomol. Chem., 2012, 10, 2318–2325
This journal is © The Royal Society of Chemistry 2012

View Online
these sites was identified as the (6–4) photoproduct.22 In other
words, the (6–4) photoproduct undergoes a reaction with a
hydroxide ion, which leads to a strand break at this site. This
property has been used to detect the formation of this type of
lesion in DNA.23–27 Terminally-labeled DNA strands were irra-
diated with UV light and treated with hot piperidine. The pro-
ducts were separated by polyacrylamide gel electrophoresis
under denaturing conditions, and bands were detected at pyrimi-
dine–pyrimidine sites. This method revealed that the frequency
of (6–4) photoproduct formation was higher at TC and CC
sites,21 although this lesion can be formed at all four types of
dipyrimidine sites.22 This alkali treatment is a well established
method, but very few studies on its reaction mechanism have
been reported. We previously found that the first reaction was the
hydrolysis of the N3–C4 bond of the 5′ component, by HPLC
and NMR analyses of alkali-treated short oligonucleotides con-
taining the (6–4) photoproduct formed at TT.28 The biochemical
properties of this hydrolyzed photoproduct were also analyzed.28
Since the ring opening at the 5′ pyrimidine site did not seem to
cause a strand break, we analyzed the following reactions occur-
ring at a higher hydroxide concentration to reveal the mechanism
of the strand break. The hydrolysis of the glycosidic bond of a
nucleoside bearing 5-methyl-2-pyrimidinone, which was used as
an analog of the 3′ component of the (6–4) photoproduct, was
also analyzed, and its mechanism was elucidated.
peak ii diminished over time, this was presumed to be the inter-
mediate (2) found in our previous study.28 The final product
(peak iii) was prepared and purified on a larger scale, and its
1
structure was analyzed by H, 13C, and 31P NMR spectroscopy,
including measurements of COSY, NOESY, HMQC, and HMBC
spectra, as shown in the ESI.† We discovered a missing set of
signals that should be assigned to one of the two sugars, and the
NOESY spectra (Fig. S3†) indicated that the lost sugar moiety
belonged to the 3′-side nucleoside. The 31P spectra revealed that
the product contained a phosphate group, and the result that the
coupling pattern of the H3′ was changed by the 31P decoupling
(Fig. S7†) indicated the position of this phosphate. These results
strongly suggested that the reaction following the intermediate
formation was the hydrolysis of the glycosidic bond at the 3′
component of the (6–4) photoproduct. This hydrolysis produces
an abasic site, which leads to a strand break by the β-elimination
reaction under alkaline conditions when the (6–4) photoproduct
resides in DNA.
Subsequently, the product yielding peak iii was analyzed by
mass spectrometry (Fig. S8†). An m/z value of 467.19, which
corresponded to [M + H]+ of the sugar-lost form, was obtained
for 3 by nano-electrospray ionization mass spectrometry, as
expected. Another strong signal was detected at the m/z value of
450.17, and this signal was observed exclusively when fast atom
bombardment was used for ionization. These results suggested
that at the ionization step in the mass analysis, a 1,3-oxazine-
2,6-dione derivative was formed at the 5′ base by a cyclization
reaction, accompanied with ammonia release. The identity of the
chemical structure of the 5′ base between 2 and the obtained
product (3) is supported by the 13C-NMR data shown in Table 1
in the ESI.†
Results
Alkali treatment of the (6–4) photoproduct (1)
We previously used a tetramer containing the (6–4) photoproduct
to detect an intermediate in the alkali degradation reactions, and
then analyzed its structure by NMR spectroscopy, using a dinu-
cleoside monophosphate in which the base moiety was the (6–4)
photoproduct. In the present study, we started with the same
dinucleoside monophosphate (1) shown in Fig. 1, to facilitate the
analysis of the chemical structure of the product. The starting
material (1) was prepared by deprotection, with ammonium
hydroxide, of the compound obtained in the course of the syn-
thesis of the oligonucleotide building block of the (6–4) photo-
product.29 This dinucleoside monophosphate was treated with 1
M NaOH at 60 °C, and the reaction mixture was analyzed by
reversed-phase HPLC (Fig. 2). Two products (peaks ii and iii in
Fig. 2b) were detected, while the same treatment of
thymidylyl(3′–5′)thymidine did not cause any reaction. Since
Fig. 2 HPLC analysis of the alkali treatment of the (6–4) photopro-
duct. Compound 1 was treated with 1 M NaOH at 60 °C for 0 h (a), 6 h
(b), or 8 h (c). The insets are UVabsorption spectra (220–365 nm) of the
major peaks, and the absorption maxima are shown.
Fig. 1 Structures of the (6–4) photoproduct (1), the intermediate found
in our previous study (2), and the product obtained in the present study
(3).
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2318–2325 | 2319

View Online
Alkali treatment of a dinucleoside monophosphate containing
5-methyl-2-pyrimidinone (4)
under alkaline conditions by the β-elimination reaction at the
resulting abasic site. This reaction was demonstrated with a
trimer containing 5-methyl-2-pyrimidinone (8), which was pre-
pared with a phosphoramidite building block.30 Alkali treatment
of this compound yielded three peaks (peaks viii, ix, and x), as
shown in Fig. 5, and the two products yielding peaks viii and x
were coeluted with authentic 7 and 5, respectively. The other
product that yielded peak ix was coeluted with 2′-deoxycytidine
5′-phosphate (9), as expected. The small peak with a retention
time of 11.3 min was 2′-deoxyuridine 5′-phosphate, which was
derived from 9 by deamination.
In order to confirm the hydrolysis of the glycosidic bond at the
3′ component of the (6–4) photoproduct and the subsequent loss
of the sugar moiety at this site, we prepared a dinucleoside
monophosphate containing 5-methyl-2-pyrimidinone30 as the
3′-side base (4 in Fig. 3), as shown in Scheme 1 in the ESI.†
This base analog resembles the 3′ component of the (6–4) photo-
product, but has no linkage between the two bases. After the
treatment of this compound with 1 M NaOH at 60 °C for 6 h,
two products (peaks v and vi) were detected by HPLC (Fig. 4).
One of the products (peak vi) was identified with thymidine
3′-phosphate (5) by a coinjection experiment (Fig. 4c), as
expected. The other product, which yielded peak v, was isolated.
This product did not have a sugar, and the H4 and the H6 were
found to be equivalent in the NMR spectrum. Although a similar
spectrum was expected for 5-methyl-2-pyrimidinone (6),31 the
UV absorption spectra (Fig. 4b) strongly suggested that the
product was 2-hydroxy-5-methylpyrimidine (7), which is a tauto-
mer of 6. This product (7) was converted to 6 by acid treatment,
as shown in Fig. S11 in the ESI.†
Alkali treatment of a trimer containing 5-methyl-2-
pyrimidinone (8)
The mechanism of the unexpected loss of the sugar moiety from
1 and 4 is described in the Discussion. On the other hand, the
hydrolysis of a glycosidic bond in DNA causes a strand break
Fig. 4 HPLC analysis of the alkali treatment of a dimer containing 5-
methyl-2-pyrimidinone. Compound 4 was treated with 1 M NaOH at
60 °C for 0 h (a) or 6 h (b and c). In panel c, thymidine 3′-phosphate
was coinjected. The insets are UV absorption spectra (220–365 nm) of
the major peaks, and the absorption maxima are shown.
Fig. 5 HPLC analysis of the alkali treatment of a trimer containing 5-
methyl-2-pyrimidinone. Compound 8 was treated with 1 M NaOH at
60 °C for 0 h (a) or 6 h (b).
Fig. 3 Structures of the compounds used in this study (4, 8, and 10)
and the products obtained by alkali treatment (5, 6, 7, and 9).
2320 | Org. Biomol. Chem., 2012, 10, 2318–2325
This journal is © The Royal Society of Chemistry 2012

View Online
Alkali treatment of a tetramer containing the (6–4)
photoproduct (10)
and xii were separately eluted in a coinjection experiment (data
not shown). As shown in Fig. 6c, this intermediate was primarily
converted into two products, which yielded peaks xiii and xiv,
after prolonged incubation. A coinjection experiment revealed
that the product that yielded peak xiii was 2′-deoxycytidine 5′-
phosphate (9). We confirmed that this type of strand break
occurred by the treatment with 0.1 M KOH at 90 °C for 30 min,
which was used to test the alkaline lability of UV-irradiated
DNA,22 but the amount of the released nucleotide was small
(Fig. S14†).
In order to confirm that the nucleotide 3′-flanking to the (6–4)
photoproduct is similarly released as a nucleoside 5′-phosphate
upon cleavage of the glycosidic bond of the 3′ component in this
photoproduct, a tetramer containing the (6–4) photoproduct (10),
which was prepared by using the oligonucleotide building
block,29 was treated with alkali. In this case, a lower concen-
tration (0.1 M) of NaOH was used to confirm the first reaction at
the 5′ base to form 2, as observed in our previous study.28 After
incubation at 60 °C for 4 h, the tetramer was completely con-
verted to this type of intermediate, as shown in Fig. 6b. Peaks xi
Comparison between the (6–4) photoproduct and its Dewar
valence isomer
The (6–4) photoproduct is isomerized to its Dewar valence
isomer by exposure to the UVA/B light,32 and this isomer is
reportedly more labile than the (6–4) photoproduct under alka-
line conditions.33 We confirmed this property by analyzing the
release of the 3′-flanking nucleotide. Another tetramer containing
the (6–4) photoproduct (11) was prepared by UV irradiation of
d(ATTG),34 and it was converted to the Dewar isomer-containing
tetramer (12) by exposure to the Pyrex-filtered light.35 Each tetra-
mer was treated with 0.1 M NaOH at 60 °C, and the products
were analyzed by HPLC at 1-hour intervals. In both cases, 2′-
deoxyguanosine 5′-phosphate was detected, and as shown in
Fig. 7, its release from 12 was much faster than that from 11. In
the case of the (6–4) photoproduct, the intermediate in which the
N3–C4 bond of the 5′ component was hydrolyzed was accumu-
lated in the same way as shown in Fig. 6b. These results indicate
that the isomerization to the Dewar isomer renders the glycosidic
bond at the 3′ component more labile.
Discussion
In this study, we investigated the mechanism of the strand breaks
at the (6–4) photoproduct sites in UV-irradiated DNA, caused by
the hot alkali treatment. In our previous study, we found that the
Fig. 6 HPLC analysis of the alkali treatment of a tetramer containing
the (6–4) photoproduct. Compound 10 was treated with 0.1 M NaOH at
60 °C for 0 h (a), 4 h (b), or 8 h (c).
Fig. 7 Comparison of the strand breaks at the (6–4) photoproduct and its Dewar valence isomer. The starting materials 11 and 12 (open circles and
open squares, respectively) and 2′-deoxyguanosine 5′-phosphate released from 11 and 12 (filled circles and filled squares, respectively) after the treat-
ment with 0.1 M NaOH at 60 °C are shown as molar ratios.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2318–2325 | 2321

View Online
first reaction in the alkali treatment of the (6–4) photoproduct
was the hydrolysis of the N3–C4 bond of the 5′ component, and
characterized the biochemical properties of this hydrolyzed
photoproduct.28 However, this reaction did not seem to lead
directly to the strand break. In the present study, the analysis of
the products obtained by treatment with a higher concentration
of NaOH revealed that the glycosidic bond was hydrolyzed at
the 3′ component of the (6–4) photoproduct in a dinucleoside
monophosphate. NMR spectroscopy and mass spectrometry ana-
lyses revealed that the 2-deoxyribose moiety at the resultant
abasic site was lost by this alkali treatment, which resulted in the
formation of 3 as a product. In DNA, β-elimination to cleave the
internucleotide linkage occurs under alkaline conditions at the
abasic site formed by hydrolysis of the glycosidic bond, but the
loss of the sugar moiety was not expected for 1, because there
was no 3′-flanking nucleotide to function as a leaving group.
When another dinucleoside monophosphate (4) containing 5-
methyl-2-pyrimidinone, an analog of the 3′ component of the
(6–4) photoproduct, was treated in the same way, the 3′-side
sugar was lost again. We propose a reaction mechanism for the
removal of the 3′-end sugar moiety, as shown in Fig. 8. To our
knowledge, this reaction has not been reported previously,
because the properties of the abasic site have been studied within
the context of oligonucleotides. We used dinucleoside mono-
phosphates to study the chemical properties of the damaged
base, but the biological significance of the reaction that occurs
only at the 3′ end, without the 3′-phosphate, may be limited.
Therefore, we analyzed the reactions using the trimer (8) and the
tetramer (10) to demonstrate the strand break by β-elimination.
The most important finding in this study was that the glycosi-
dic bond at the 3′ component of the (6–4) photoproduct was
hydrolyzed by the hot alkali treatment, because this hydrolysis
causes a strand break at this site by the β-elimination reaction, as
shown in Fig. 9. It was shown that the strand break by the same
mechanism occurred in a tetramer containing the Dewar valence
isomer of this photoproduct at a reaction rate higher than that
obtained for the (6–4) precursor (Fig. 7). This observation agrees
with the previous reports showing higher alkaline lability of the
Dewar isomer.33,36,37 When we tested the conditions of 0.1 M
KOH at 90 °C for 30 min, which were used to detect the (6–4)
photoproduct sites in the past,22 the strand break was not com-
pleted (Fig. S14†). These results strongly suggest that the strand
Fig. 8 Proposed mechanism for the loss of the 3′ sugar after hydrolysis of the glycosidic bond.
Fig. 9 Mechanism of alkali degradation at the (6–4) photoproduct site.
2322 | Org. Biomol. Chem., 2012, 10, 2318–2325 This journal is © The Royal Society of Chemistry 2012

View Online
breaks detected under such conditions were caused mainly at the
sites of the Dewar photoproduct. However, it should be noted
that the (6–4) photoproduct is not intact after this treatment
because the hydrolytic ring opening of its 5′ component28 occurs
even at the lower alkali concentration, as shown in Fig. S14.†.
We used 5-methyl-2-pyrimidinone as a model compound for
the 3′ component of the (6–4) photoproduct. When this base
analog is incorporated into oligonucleotides, its glycosidic bond
is reportedly labile under both acidic38 and alkaline39 conditions,
although the mechanism has not been investigated. In our
present study, 2-hydroxy-5-methylpyrimidine (7), a tautomer of
5-methyl-2-pyrimidinone (6), was obtained by the hot alkali
treatment. The tautomerism between 2-hydroxypyrimidine and
2-pyrimidinone has been studied mainly by theoretical
methods.40–42 The two tautomers can be distinguished easily by
UV-absorption, whereas two singlet signals are observed for
both of the tautomers by 1H-NMR spectroscopy, due to the sym-
metry of these molecules. In this study, the enol form, with a UV
absorption maximum at a shorter wavelength, was obtained as
an isolated product (7), and was also detected immediately after
the reaction (peak v in Fig. 4b). In our previous study, acid
hydrolysis of 1-(2-deoxy-β-D-ribofuranosyl)-5-methyl-2-pyrimi-
dinone yielded 6, which had an absorption maximum at
311 nm.14 These results indicate that the acid and alkali hydroly-
ses of the glycosidic bond of this modified base exclusively
produce the keto and enol forms, respectively. This observation
corresponds to each reaction mechanism. In an acidic solution,
protonation occurs at N3,14 which induces glycosidic bond clea-
vage in an SN1 reaction with a water molecule at the C1′ pos-
ition. By contrast, the attack of ammonia at C6, leading to
saturation of the C5–C6 bond, was proposed as the mechanism
of glycosidic bond cleavage during the deprotection of synthetic
oligonucleotides containing this base analog,39 and in our study,
hydration products were detected by LC-MS after the hot alkali
treatment, as shown in Fig. S12 in the ESI.† Since the resonance
stabilization of the base moiety is lost, due to the saturation of
the C5–C6 bond, an SN2 reaction with a hydroxide ion occurs at
C1′. In this case, the enol form is produced by the displacement
of the electron pair from the double bond in the C2 carbonyl
group to the oxygen atom, and the aromatic pyrimidine ring is
reproduced by the following dehydration (Fig. S13†). Although
the mechanism could be the same, there was a difference in the
product obtained after the glycosidic bond hydrolysis between
5-methyl-2-pyrimidinone and the (6–4) photoproduct. The
product obtained from the compound containing the (6–4)
photoproduct had an absorption maximum at 317 nm, as shown
in Fig. 2b, which indicated that the structure of the 3′ base
released from the sugar moiety was the keto form (Fig. 9). The
5′ base linked at the C4 position may have some influence on the
tautomerism of the 3′ component.
β-elimination reaction occurs at the resultant abasic site. The
mechanism for the alkali hydrolysis of the glycosidic bond of
1-(2-deoxy-β-D-ribofuranosyl)-5-methyl-2-pyrimidinone, which
resembles the 3′ component of the (6–4) photoproduct, was also
elucidated. The chemical basis of the alkali lability of the (6–4)
photoproduct-containing DNA obtained in this study will con-
tribute to molecular and cellular biology. For example, the (6–4)
photoproduct in cellular DNA may be degraded artificially, by
targeting the glycosidic bond of the 3′ component.
Experimental
Materials
All solvents and reagents were obtained from Wako Pure Chemi-
cal Industries, except for 2′-deoxycytidine 5′-phosphate, which
was obtained from Sigma. Reagents for the DNA synthesizer
(Applied Biosystems 3400) were purchased from Glen Research.
For column chromatography, Wakogel C-200 (Wako Pure
Chemical Industries) was used. Cation exchange was performed
using AG 50W-X2 resin (Bio-Rad Laboratories).
HPLC analysis and purification
HPLC analysis was performed on a Gilson gradient-type analyti-
cal system equipped with a Waters 2996 photodiode array detec-
tor. HPLC purification was performed on the same system
equipped with a Gilson 151 UV/Vis detector. An Inertsil ODS-3
5 μm column (4.6 × 250 mm, GL Sciences) was used at a flow
rate of 1.0 ml min−1 for analysis, and a μ-Bondasphere C18
15 μm 300A column (300 × 7.8 mm, Waters) was used at a flow
rate of 2.0 ml min−1 for purification. A linear gradient of aceto-
nitrile in 0.1 M triethylammonium acetate (pH 7.0) was used for
both analysis and purification.
NMR spectroscopy
1H NMR spectra were measured at 30 °C on a Varian Unity-
1
INOVA 500 or JEOL JNM-AL400 spectrometer. The H chemi-
cal shift was calibrated with internal TMS (0 ppm) in deuterated
organic solvents or with HDO (4.70 ppm) in D2O. 13C and 31P
NMR spectra were measured on a Varian Unity-INOVA 500
spectrometer. In the 13C measurement in D2O, the 13C reference
frequency was obtained by calculation from the 1H reference fre-
quency, as reported previously.43 Briefly, after adjustment of the
1
1H chemical shift, the resonance frequency at 0 ppm for H was
converted to that for 13C by multiplying by the frequency ratio
of 13C/1H. The 31P chemical shift was calibrated with external
trimethyl phosphate. Two-dimensional NMR spectra were
recorded on a 5 mm pulse field gradient probe for indirect detec-
tion on the Varian Unity-INOVA 500 spectrometer. For the
NOESY measurement, the mixing time was set to 700 ms.
Conclusion
We have elucidated the mechanism for the strand breaks caused
at (6–4) photoproduct sites in UV-irradiated DNA by the hot
alkali treatment. Independently of the ring opening at the 5′ com-
ponent, which we reported previously, the glycosidic bond is
cleaved at the 3′ component, and strand breakage by the
Mass spectrometry
Mass spectra were acquired on a JEOL JMS-700 mass spec-
trometer. The LC-MS analysis of the alkali-treated sample of
1-(2-deoxy-β-D-ribofuranosyl)-5-methyl-2-pyrimidinone (11 in
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2318–2325 | 2323

View Online
Alkali treatment of a tetramer containing the Dewar valence
isomer of the (6–4) photoproduct (12)
ESI†) was performed with an Agilent 1200SL HPLC system
equipped with a Micromass Q-Tof Premier mass spectrometer.
The electrospray interface was operated in the ESI negative ion
mode. The capillary voltage was maintained at 2.5 kV, and the
voltages of the sample cone and the collision were set to 20 V
and 4.0 V, respectively. The elution conditions were the same as
described above.
Compound 12 (20 nmol) and thymine (20 nmol) were dissolved
in 0.1 M NaOH (100 μl), and the mixture was incubated at
60 °C. At 1-hour intervals, 10 μl of the solution was neutralized
with 0.1 M HCl, and was analyzed by reversed-phase HPLC,
using a linear gradient 0–25% acetonitrile. The release of
2′-deoxyguanosine 5′-phosphate was confirmed by a coinjection
experiment. For comparison, the alkali degradation of 11 was
analyzed in the same way. From the peak areas, molar ratios
were calculated using thymine as an internal standard.
Alkali treatment of the (6–4) photoproduct (1)
Compound 1 (128 μmol) was dissolved in 1 M NaOH (1.0 ml)
and incubated at 60 °C. After neutralization with 1 M HCl, the
mixture was analyzed by reversed-phase HPLC using a linear
gradient of 0–10% acetonitrile. The product obtained by an 8 h
alkali treatment of 1 was purified by reversed-phase HPLC with
a linear gradient of 0–5% acetonitrile during 20 min. The cation
was then exchanged for Na+. The structure of this product was
analyzed by NMR spectroscopy and mass spectrometry. 1H
NMR (500 MHz, D2O): δ (ppm) 7.99 (s, 1H, pT-H6), 5.85 (dd,
J = 9.9, 5.1 Hz, 1H, H1′), 5.51 (s, 1H, TpH6), 4.42 (m, 1H,
H3′), 4.04 (m, 1H, H4′), 3.65 (m, 2H, H5′), 2.23 (s, 3H, pT-Me),
1.98 (m, 1H, H2′), 1.50 (m, 1H, H2′), 1.24 (s, 3H, Tp-Me). 13C
NMR (125 MHz, D2O): δ (ppm) 179.9 (TpC4), 174.8 (pTC4),
161.8 (pTC2), 161.5 (TpC2), 152.7 (pTC6), 117.3 (pTC5), 87.8
(C4′), 87.0 (TpC5), 86.7 (C1′), 77.1 (C3′), 65.0 (C5′), 61.9
(TpC6), 39.4 (C2′), 21.1 (TpMe), 15.9 (pTMe). 31P NMR
(203 MHz, D2O): δ (ppm) 4.06. ESI-LRMS m/z: 467.19 ([M + H]+
calcd. for C15H24N4O11P: 467.12).
References
1 G. C. Walker, W. Siede, R. A. Schultz, T. Ellenberger and
E. C. Friedberg, in DNA Repair and Mutagenesis 2nd Edn, ASM Press,
Washington D. C., 2006.
2 S. Iwai, in Modified Nucleosides in Biochemistry, Biotechnology and
Medicine, ed. P. Herdewijn, Wiley-VCH, Weinheim, 2008, pp. 97–131.
3 S. Y. Wang and A. J. Varghese, Biochem. Biophys. Res. Commun., 1967,
29, 543–549.
4 A. J. Varghese and S. Y. Wang, Science, 1968, 160, 186–187.
5 I. L. Karle, S. Y. Wang and A. J. Varghese, Science, 1969, 164, 183–184.
6 R. O. Rahn and J. L. Hosszu, Photochem. Photobiol., 1969, 10, 131–
137.
7 J. E. LeClerc, A. Borden and C. W. Lawrence, Proc. Natl. Acad.
Sci. U. S. A., 1991, 88, 9685–9689.
8 C. A. Smith, M. Wang, N. Jiang, L. Che, X. Zhao and J.-S. Taylor, Bio-
chemistry, 1996, 35, 4146–4154.
9 A. Gentil, F. Le Page, A. Margot, C. W. Lawrence, A. Borden and
A. Sarasin, Nucleic Acids Res., 1996, 24, 1837–1840.
10 H. Kamiya, S. Iwai and H. Kasai, Nucleic Acids Res., 1998, 26, 2611–
2617.
11 M. J. Horsfall and C. W. Lawrence, J. Mol. Biol., 1994, 235, 465–471.
12 C. Masutani, M. Araki, A. Yamada, R. Kusumoto, T. Nogimori,
T. Maekawa, S. Iwai and F. Hanaoka, EMBO J., 1999, 18, 3491–3501.
13 C. Masutani, R. Kusumoto, S. Iwai and F. Hanaoka, EMBO J., 2000, 19,
3100–3109.
14 J. Yamamoto, Y. Tanaka and S. Iwai, Org. Biomol. Chem., 2009, 7, 161–
166.
15 K. Sugasawa, Mutat. Res., Fundam. Mol. Mech. Mutagen., 2010, 685,
29–37.
16 T. Nouspikel, Cell. Mol. Life Sci., 2009, 66, 994–1009.
17 J. Li, Z. Liu, C. Tan, X. Guo, L. Wang, A. Sancar and D. Zhong, Nature,
2010, 466, 887–890.
18 Z. B. Yang, R. B. Zhang and L. A. Eriksson, Phys. Chem. Chem. Phys.,
2011, 13, 8961–8966.
19 Y.-J. Ai, R.-Z. Liao, S.-F. Chen, Y. Luo and W.-H. Fang, J. Phys. Chem.
B, 2010, 114, 14096–14102.
20 K. Sadeghian, M. Bocola, T. Merz and M. Schütz, J. Am. Chem. Soc.,
2010, 132, 16285–16295.
Alkali treatment of a dinucleoside monophosphate containing
5-methyl-2-pyrimidinone (4)
Compound 4 (1 μmol) was dissolved in 1 M NaOH (1.0 ml) and
incubated at 60 °C. After neutralization with 1 M HCl, the
mixture was analyzed by reversed-phase HPLC, using a linear
gradient of 0–25% acetonitrile. Thymidine 3′-phosphate, which
was synthesized on a DNA synthesizer using 5′-O-(4,4′-
dimethoxytrityl)thymidine 3′-(2-cyanoethyl)-N,N-diisopropyl-
phosphoramidite and 3′-phosphate CPG (Glen Research), was
then coinjected into an HPLC column along with the sample
treated with hot alkali for 6 h. The base moiety released by the
alkali treatment was purified by reversed-phase HPLC with a
linear gradient of 0–10% acetonitrile during 20 min. After the
cation exchange for Na+, the structure of this product was ana-
lyzed by NMR spectroscopy. 1H NMR (400 MHz, D2O): δ
(ppm) 8.39 (s, 2H, H4 and H5), 1.51(s, 3H, Me).
21 J. A. Lippke, L. K. Gordon, D. E. Brash and W. A. Haseltine, Proc. Natl.
Acad. Sci. U. S. A., 1981, 78, 3388–3392.
22 W. A. Franklin, K. M. Lo and W. A. Haseltine, J. Biol. Chem., 1982,
257, 13535–13543.
23 D. E. Brash and W. A. Haseltine, Nature, 1982, 298, 189–192.
24 F. Bourre, G. Renault and A. Sarasin, Nucleic Acids Res., 1987, 15,
8861–8875.
Alkali treatment of a trimer containing 5-methyl-2-
pyrimidinone (8)
25 G. P. Pfeifer, R. Drouin, A. D. Riggs and G. P. Holmquist, Proc. Natl.
Acad. Sci. U. S. A., 1991, 88, 1374–1378.
26 J. Jen, D. L. Mitchell, R. P. Cunningham, C. A. Smith, J.-S. Taylor and J.
E. Cleaver, Photochem. Photobiol., 1997, 65, 323–329.
27 J.-H. Yoon, C.-S. Lee, T. R. O’Connor, A. Yasui and G. P. Pfeifer, J. Mol.
Biol., 2000, 299, 681–693.
28 M. Higurashi, T. Ohtsuki, A. Inase, R. Kusumoto, C. Masutani,
F. Hanaoka and S. Iwai, J. Biol. Chem., 2003, 278, 51968–51973.
29 S. Iwai, M. Shimizu, H. Kamiya and E. Ohtsuka, J. Am. Chem. Soc.,
1996, 118, 7642–7643.
30 S. F. Singleton, F. Shan, M. W. Kanan, C. M. McIntosh, C. J. Stearman,
J. S. Helm and K. J. Webb, Org. Lett., 2001, 3, 3919–3922.
Compound 8 (500 nmol) was dissolved in 1 M NaOH (1.0 ml)
and incubated at 60 °C. After neutralization with 1 M HCl, the
mixture was analyzed by reversed-phase HPLC, using a linear
gradient of 0–25% acetonitrile. Thymidine 3′-phosphate and
2′-deoxycytidine 5′-phosphate were then coinjected into an
HPLC column along with the sample treated with hot alkali
for 6 h.
2324 | Org. Biomol. Chem., 2012, 10, 2318–2325
This journal is © The Royal Society of Chemistry 2012

View Online
31 S. Buchini and C. J. Leumann, Eur. J. Org. Chem., 2006, 2006, 3152–3168.
32 J.-S. Taylor and M. P. Cohrs, J. Am. Chem. Soc., 1987, 109, 2834–2835.
33 L.-S. Kan, L. Voituriez and J. Cadet, J. Photochem. Photobiol., B, 1992,
12, 339–357.
38 J. A. Iocono, B. Gildea and L. W. McLaughlin, Tetrahedron Lett., 1990,
31, 175–178.
39 B. A. Connolly and P. C. Newman, Nucleic Acids Res., 1989, 17, 4957–
4974.
34 C. A. Smith and J.-S. Taylor, J. Biol. Chem., 1993, 268, 11143–11151.
35 J.-S. Taylor, D. S. Garrett and M. P. Cohrs, Biochemistry, 1988, 27,
7206–7215.
36 D. L. Mitchell, Mutat. Res. DNA Repair, 1988, 194, 227–237.
37 T. Douki, L. Voituriez and J. Cadet, Photochem. Photobiol., 1991, 53,
293–297.
40 C. Ögretir and M. Yaman, THEOCHEM, 1999, 458, 217–226.
41 C. W. Bird, Tetrahedron, 1992, 48, 7857–7862.
42 A. Les and L. Adamowicz, J. Phys. Chem., 1990, 94, 7021–7032.
43 D. S. Wishart, C. G. Bigam, J. Yao, F. Abildgaard, H. J. Dyson,
E. Oldfield, J. L. Markley and B. D. Sykes, J. Biomol. NMR, 1995, 6,
135–140.
This journal is © The Royal Society of Chemistry 2012
Org. Biomol. Chem., 2012, 10, 2318–2325 | 2325